Peanut Tin of Terror

Recent Construction Details

Because I was working on the Peanut Tin of Terror a lot before Combots I didn�t put up too many pictures of the work I was doing. I thought I should catch up a little.

Here you can see the new blade for the Peanut Tin�s Weapon. To make this shape I drew the pattern on a piece of paper with a ruler. Then I stuck the paper to a piece of sheet glue with some rubber cement. I cut the sheet steel with a jig saw and then removed the paper pattern. Then I painted the steel black. The sharp corners give the spinning surface teeth.

Here you can see the color sensors that I put in front of the robot. Originally I planned to put them inside the peanut tin but with that configuration they did not give enough warning of the oncoming white stripe. Where the original hole was drilled (near the center of the peanut tin as you can see below) by the time the detector sensed white there was a good chance that one of the two wheels was already over the edge and the robot was immobile.

Ultimately peanut tin lost at Combots because it went out of the ring, usually on its own. John Frizell is very nice and sent me a version of one of his line detectors after our competition, I think because he thought this was where my robot failed. I think the problem was more software related because I wrote the original program for a ring twice as large as the one we ended up using. In the larger ring there was more room to back up-- in other words you could drive into the edge ring and very safely back up two feet. In the smaller ring at certain angles doing that would cause the robot to back over the edge. I think I can do a lot better by just taking out the backing up part.

Peanut Tin at Combots V

The Peanut Tin of Terror was originally created to battle at Combots IV. Unfortunately I was unable to attend that competition and so Peanut had its first battle at Robogames. This year I am looking forward to Combots.

I have made a lot of improvements to the robot. I moved the sonar sensors down and to the sides of the blade so that they could detect low robots. I added color sensors to the bottom front of the robot so that I could detect white lines at the edge of the combat arena.

I also tried a cool new site called Robotbox. You can comment or make suggestions at the site for the Peanut Tin of Terror.

Peanut Tin at Robogames 2010

Here is the peanut tin in action, just before Robogames 2010.

Peanut tin won second place in autnonomous combat.

December 21st Problems with H-Bridge

I am having trouble getting the H-Bridge to work on the Parallax robot controller board. I can always switch to a Sabertooth RC-5 speed controller if I have to, but I hope I can get this to work because the new board is cool.

I am linking this to a question thread in the Parallax robotics forum, so please forgive the boring pictures and movies of what I am doing. I am hoping someone will see this here and be able to help me by making suggestions on the forum.

Here is a picture of the insides of the peanut tin.

Here is a picture of the controller board hooked up, plugged in, and turned on.

December 6th - Sensor Tower

This weekend I was able to complete the sensor tower for the peanut tin. I think this is a pretty good array of sensors. Since the peanut tin isn't going to be tougher, I hope it will have the senses it needs to be a little smarter. There are two sonar sensors and a motion detector.

The sensors are raised in order to avoid the rotating weapon in the front. Normally I try to keep obstacle sonar at about as low of a horizontal height as I can to detect low obstacles. I didn't want to put them to the sides, because that makes them vulnerable. Also, I think I can take advantage of the special environment of the battle box.

The battle box has high, flat walls. There are no low obstacles other than the other robot. Therefore for collision avoidance detecting walls at a higher point then the robot's wheels is OK. The sensors are likely to overshoot an opponent, but they are not meant to detect other robots. My plan is to use the motion detector to find the enemy robot. We will see how well that works.

The tower is made from 0.025” aluminum, cut with a saw and tin snips. It is very light, but it may wobble too much to get reliable sensor readings.

December 2nd - Cool New Parts

I got some cool new parts for the Peanut Tin. For power I got three 7.2v 1600mAh batteries from Tower Hobbies. They are light-- less than 6 ounces-- and seem to have adequate power for a three minute match. One 7.2v battery will run the servos and the microcontroller. The other two will run the weapon motor.

Since I started Peanut Tin Parallax has started selling a robot controller board. I have been waiting for a propeller board made for robots so long now that I can't resist putting it in. The new board has a pair of built in H bridges. If I use one of these for the weapon motor I will not need to add an electronic speed controller. An ESC works a little better than an H-Bridge, but both will probably work well enough.

What's the green part in the picture? That's an interesting sensor that I have never tried to use before. I will explain more as I test it out.

November 21st - Progress and Batteries

I have been making steady progress on the Peanut Tin, but because of my involvement with the PTA-- don't ask-- I have been too busy to post. Most of the construction is coming together, except for mounting sensors which I expect to be the last task.

Now that I have a clearer picture of the overall weight, I need to decide what sort of batteries to use. The question I face in particular is whether to switch to lithium polymer (LiPo) batteries. So far all the robots I have built use Nickel Cadmium (NiCd) or Nickel Metal Hydride (NiMH) batteries. NiCd/ NiMH can use the same sort of charger, which I already own. LiPo batteries have better power to weight than NiMH or NiCD, I am guessing 33%- 66% better depending on what you are comparing it to. But LiPo batteries require a special charger that seems to run about $120 dollars. They are also supposed to be tricky to charge and to balance. The average LiPo battery is more expensive. The premium is decreasing, but a LiPo battery will probably always be more than a NiMH/ NiCD battery. For now I have decided to go with three small 7.2v NiMH batteries, which will run me about $65. Eventually I am going to have to switch to LiPo.

In these pictures you can see I mounted a metal strip on the gear motor as a weapon. This was a very simple thing, I just cut a piece of sheet aluminum I got from the hardware store. Unfortunately it probably isn't a very good solution. Aluminum is softer than steel and lighter. Light is usually good, but since my front spinner is only doing 300-400 rpm I probably want it to have some weight so that it does damage when it hits. I could add weights to this aluminum sheet, or better still spikes. The weapon is an easy component to replace, so it makes sense to get something working and then replace it later. I am worried that the motor I am using is just too slow to be dangerous. That would be tougher to replace so I hope I can make it work.

October 31st - Parts Arrive

Last week I got a protoboard from a parallax. This Monday I got the parts I ordered from Servocity. I ordered more than just the parts for the Combot, and I did the cheapest shipping possible. I opened the box from Servocity with my six year old son. There was some candy inside, I guess for Halloween. Talk about a treat for everyone! Anyone, here are the fixings for my Combot robot:

October 31st - Peanut Tin of Terror

With the parts here I have begun construction. You can see the robot start to take shape. I have decided to name it after the can I am using for a chassis. If you have never had hubs peanuts, they are delicious. My father's company gives them out as Christmas presents every year to all of their clients. I have really come to love them.

Planning for Combots

I spent some time this weekend thinking about what the robot I enter for combots could be like. Why am I entering the autonomous 3lb competition? I am terrible at construction, so I don't think I would be very good at radio controlled robot combat. Autonomous combat might give me more of a chance because of the difficulty of creating self-controlling robots. My goal will be to make a robot that isn't necessarily faster or more durable. It will have a decent attack and be smart enough to strike.

Planning how much to spend on parts for this robot is quite difficult since it may be destroyed. How much do you want to spend on a robot that will probably be destroyed? Ideally many of the components will survive, but you don't want to assume they will.

Using gear motors for the drive train is preferable for performance, but it would add the requirement of having a speed controller for each wheel. A speed controller is a $50 component. Gear motors are also a little more expensive then servos. So I decided to go with full rotation servos, even though that probably really limits the power of my robot.

For sensors I would like to use the Ping sonar sensor from Parallax. I have used this a ton and it works. I have used it a lot on Fluffy, you can see how it works in navigation in videos. With sensors I think it makes sense to go with what you know, the hardest part is getting them to work reliably.

For a controller I will stick with the Propeller chip. I have been using them so much lately it makes sense to stick with them for this application. However, instead of using a Spin Studio card as I usually do, I think I will use a more simple propeller protoboard from parallax. That board is lighter and cheaper. Lighter is important because of the weight limit. Cheaper is good because it is apt to get destroyed.

For a weapon I plan to use a high RPM gear motor. A spinning weapon is simple and likely to be effective against another robot carrying a micro controller.

How does this plan break out into parts? Here is my list: